Holy crap! HUGE diff in Core Temp beta! Ack!

Toggle sidebar Toggle sidebar
E

EricVPI

Distinguished
May 15, 2006
166
0
18,680
All this time I have been using Core temp regular version to measure my temps.

I tried the beta version AND MY TEMPS HAVE LITERALLY INCREASED 14C. Which one is real? Probably the beta?


I'm at STOCK and I'm running 68C max load. That CANNOT be right. According to old coretemp I was at 54C full load.

Something is wrong.


I have an e6850.
 
cnumartyr

cnumartyr

Distinguished
Nov 3, 2007
2,287
0
19,780
It's right ;)

Old Core Temp reads the Tjunction Max as 85 C.

G0 cores have a Tjunction Max of 100 C.

Because of that the old one would read 15 C low on the new G0s.

The good news is you are within Intel specs. 71 C is the Tcase. Read the guide in the OCing section on Temps and you will see that up to 75 C is fine with what you have for Core Temps.
 
E

EricVPI

Distinguished
May 15, 2006
166
0
18,680
Yeah, but I'm only at stock. There goes my hope of overclocking to 3.6, HA.

Maybe I will have to cough up for that aftermarket cooler. But I don't want to unplug all my mobo stuff and unscrew the board and undo the heatsink and scrape off the gum and...


(ick)
 
cnumartyr

cnumartyr

Distinguished
Nov 3, 2007
2,287
0
19,780


I've done that 6 times already in 2 weeks. I reseat it over and over til I'm satisfied. Still gotta do it again for MX-2. :p
 
E

EricVPI

Distinguished
May 15, 2006
166
0
18,680
On the first computer I built years ago I messed up when seating the mobo against the standoffs and the mobo just stopped working.

Ever since then I have had a chronic fear of seating mobos, much less taking off HSF.
 
cnumartyr

cnumartyr

Distinguished
Nov 3, 2007
2,287
0
19,780
I burned my first build by forgetting the stand offs at all. Burned the PSU up.

I burned one the other week because I got AS5 on the mobo and never wiped it off.
 
nukemaster

nukemaster

Titan
Nov 23, 2004
19,626
6
74,965
Try to reseat the heat sink. And if you have any kind of fan control turn it off...

LOAD(10 min of folding)
loadtg0.gif

IDLE(doing nothing for 10 min)
idleko8.gif

This NOT the stock heatsink tho....I put that on my friends E6750 :)
 
nukemaster

nukemaster

Titan
Nov 23, 2004
19,626
6
74,965


No its a hotter(quad) proc with a low speed fan and identical clock. My point is something is mounted wrong....want me to post with the stock heatsink? I did just setup one with the stock heat sink and over clocks...

If i wanted to brag i would crank the fans and up it to 3.6 so "BLAHHH" next time....
 
cnumartyr

cnumartyr

Distinguished
Nov 3, 2007
2,287
0
19,780


I agree it may be mounted wrong. But you definately don't prove it by posting with a different processor, different heat sink, and different everything. :kaola:
 
nukemaster

nukemaster

Titan
Nov 23, 2004
19,626
6
74,965
thats all the point was....had nothing to do with any kind of brag(and i take back any mentioned stfu'ing)....but i ran a stock unit(i have no access to it....one i built for a friend who lives far enough away..)and it ran very cool. Cooler then my E6600 system ever did.....i was surprised to see a G0 run that hot....
 
Grimmy

Grimmy

Splendid
Feb 20, 2006
4,431
0
22,780


You mention older CoreTemp? Was it the very first one, version .94?

Because when .95 came out, it read hotter readings, and this was listed as one of their fixes when .95 came out:

- Fix: Core Temp reported wrong temps after a while. (Driver conflict)
Click to expand...

The only possiblity I see, is either poor airflow, or the stock HS doesn't have all 4 pins secured to the MB.
 
K

kuniskos

Distinguished
Mar 20, 2006
37
0
18,530
I'm new to overclocking and also have an e6850, which I've been fooling with for a while now. This is the nutshell version.

EVGA A1 680i, Sonata III, EarthWatts 500w PSU, 2 x BFG 7900GS, 2 SATA HDs, 1 SATA DVDRW, 2GB Corsair DDR2 800 (5-5-5-18), Zalman 9700.

The only OC I've been happy with is 3.6 GHz. I managed 3.825 Prime95 stable @ 1.45v but had very high temps. Temps at 3.6 GHz are still high but performance is good, and after reading CompuTronix & graysky's instructions on temps & overclocking, I'm okay, for now, risking the increased heat. The Sonata III was probably not the best choice but I'm reluctant to make a change now. (Though I did just see a Cooler Master CM 690 for $40 after MIR, and it would probably be better.)

The Zalman 9700 + Arctic Silver 5 made a big difference. My first HSF change was to an Arctic Cooling Freezer 7 Pro (with included thermal material), but the improvement was minimal so I sprang for the 9700. The installation was worth the trouble.

I don't think CPU temp (Tcase) has ever been accurate (always right around or a little above core temps), but I haven't yet done the calibration detailed by CompuTronix.

I watch core temps mostly and am (for now anyway) willing to take whatever risk is associated with generally running in the mid to high 40s or 50s. A lot of temps people offer up here are lower than actual, I'm guessing, but I also know mine aren't great. On the other hand, I'm never hitting the high 70s (and whatever the true Tcase might be) that occur during a 100% TAT run.

Using Auto in the BIOS never worked out very well because it seemed pretty clear that some voltages were being set way to high.

(At 3.0 GHz (stock), load temps weren't that much lower than what follows.)

Values:

Stable 680i BIOS voltages (3.6 GHz), all set manually:
CPU: 1.3875
FSB: 1.3
RAM: 1.85
PCI-E: 1.3
NForce/AUX: 1.5
SPP-MCP: 1.2

3.6 GHz Temps Case Fully Closed (read by Core Temp 95.4 & SpeedFan/NVIDIA Monitor):
Idle
(core1/core2/cpu)
46/45/49
vcore = 1.368 actual (1.3875 in bios)

load (Prime95 v. 25.4)
(core1/core2/cpu)
63/63/66
vcore = 1.344 actual (1.3875 in bios)

load (Intel TAT 100% both cores)
(core1/core2/cpu)
79/77/78
vcore = 1.328 actual (1.3875 in bios)
 
nukemaster

nukemaster

Titan
Nov 23, 2004
19,626
6
74,965
you can gain some performance from your Sonata III by opening the front door and taking out a CD slot cover(the case does not have enough vents to let air in....and in fact starves that rear fan...low is ok....med starts to starve and high does not give the fan enough air....)....Sonata III is more of a stock case....not so good for OCing. The 900 has plenty of airflow for such overclocking.
 
K

kuniskos

Distinguished
Mar 20, 2006
37
0
18,530
Thanks very much, nukemaster. Great idea. And easy. I removed 2 slot covers, adjusted the fans very slightly, and cores dropped from 47/45 to 42/39. Tcase dropped 4 or 5C too. After some fooling around I'm convinced these numbers are stable and attributable to your fix.

You really nailed it, and the great thing is that I don't even notice additional noise. The only problem is that I now want to go back to 3.825.
 
CompuTronix

CompuTronix

Intel Master
Moderator
Nov 13, 2006
6,374
279
38,240
kuniskos, it's very tempting to crank it up higher, however, you may first want to perform the Calibrations in the Guide to offset your inaccurate Tcase, and to make certain that your temps are correct. The procedure is simple enough, and it works.

Comp :sol:
 
K

kuniskos

Distinguished
Mar 20, 2006
37
0
18,530
Thanks, Comp. I'd say you're right: now is definitely the time for the calibrations.

Here's a little more detail following nukemaster's advice (with a high Tcase reading after a few minutes of TAT):

Idle
(core1/core2/cpu)
42/39/45

load (Prime95 v. 25.4)
(core1/core2/cpu)
56/56/59

load (Intel TAT 100% both cores)
(core1/core2/cpu)
72/71/72
 
CompuTronix

CompuTronix

Intel Master
Moderator
Nov 13, 2006
6,374
279
38,240
kuniskos, I'm puzzled by your Core (Tjunction) temperatures. The difference between TAT and Prime95 should only be about 5c, IF you're running Prime95 as intended, (Small FFT's for CPU Load testing), which is Prime95's heaviest processor Load, and is the same as Orthos. Remember that TAT's 100% Thermal Load (all one's through all registers), would be equivalent to Prime95 Small FFT's at 114% Work Load. TAT is the the most extreme CPU thermal testing available.

TAT exceeds Prime95 or Othos by 14%, and far exceeds any real-world video processing or gaming loads that would ever be applied to your processor, which is also true for Prime95 and Orthos. The most processor intensive apps and games running simultaneously would rarely reach a sustained 80% Work Load, so with this in mind, TAT represents an unrealistic CPU thermal test, except that it's useful to establish the extreme upper limits of processor and computer case cooling capacity.

So my question is, are you running Prime95's Torture Test using Small FFT's?

Comp :sol:
 
K

kuniskos

Distinguished
Mar 20, 2006
37
0
18,530
Thanks again, Comp. You're exactly right. I was running a blend, and getting a lot of variance, which really is, and was, misleading. The temps below, using Small FFTs, are more telling.

The load percentage info is very helpful too.

I actually just started a new thread on the calibrations, which I finally did though I had some trouble setting my CPU to 1.6 GHz (couldn't even do it). But I think the final Tcase offset is about right.

EricVPI's many posts, and the many great replies, got me looking into all this in more detail. Glad I did, and thankful for all your help.

New Temps - all P95 values refer to Small FFTs - @3.735 GHz:

load (Prime95 v. 25.4, Small FFTs)
(core1/core2/cpu)
70/70/65

load (Intel TAT 100% both cores)
(core1/core2/cpu)
77/74/67

idle (but recently active):
(core1/core2/cpu)
43/42/37
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom